0

I want to use GAM method and generalized exponential distribution for response variable. I know GAM method is a generalized GLM method and the distribution of response variable must be in exponential family. The probability density (pdf) of generalized exponential distribution is as following :

$$ f(x ; \alpha, \eta)=\alpha \eta \exp\left\{ -\eta x \right\}\cdot \left( 1-\exp(-\eta x) \right)^{\alpha - 1}, \quad x>0 $$ CDF of this distribution is as following : $$ F(x; \alpha, \eta) = \left(1-\exp(-\eta x)\right)^\alpha, \quad x>0 $$ The $\alpha$ is shape parameter and the $\eta# is scale parameter. How can I write this pdf as exponential family? That is, is the generalized exponential distribution a member of the exponential family? Also known as the *exponentiated exponential distribution, a special case of the Exponentiated Weibull distribution

kjetil b halvorsen
  • 77,844
user321525
  • 71
  • At https://stats.stackexchange.com/a/519715/919 I describe a general method to solve this problem -- maybe it will help you with this one. – whuber May 12 '21 at 19:44
  • Thank you @Whuber I have two parameters. For using this method, should I differentiate first w.r.t x and then w.r.t alpha and finally w.r.t eta? – user321525 May 13 '21 at 05:36

1 Answers1

0

No, you cannot. Just take the logarithm of the density, and you will see that the expression do not take the form of the argument to the exponential function needed for an exponential family. You need that the interaction of the argument $x$ and the parameters $\theta$ can be separated as $ \eta(\theta) \cdot T(x)$, where the dot represents a dot product. You don´t get that form.

But you want this for a gam, maybe look into generalized non-linear models, see how to fit a generalized nonlinear model in R? or https://stackoverflow.com/questions/59493784/writing-a-custom-model-for-gnm-package-but-need-a-function-not-in-derivative-ta

kjetil b halvorsen
  • 77,844